This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

DAC8750: DAC8750 retrofit into Analog Devices AD5420 Design

Part Number: DAC8750
Other Parts Discussed in Thread: DAC8760

We are evaluating the feasibility of switching from the Analog Devices' AD5420 to the TI DAC8750. Initial results have been promising: we soldered an DAC8750 onto our board and had the correct current output with no software modifications.

Our main concern is the DAC8750 capability of turning on the watchdog and CRC mode without a LATCH signal (Tables 2 and 4 of the datasheet).  We use a party-line SPI bus, so unlatched command words are a major downside.

To compensate, we are considering turning on the watchdog and CRC immediately after boot up, then we do not have to worry about a ``lucky'' SPI message turning these unlatched features on.

My question: We have had EMI problems with the AD5420 where the output range shifts from a normal 4-20mA to a crazy output of (in one case) 28mA to 134mA.

Is there any public EMI test data available for the DAC8750 that would give us assurance that it is better in this respect than the AD5420 ?


  • Hi Tom,

    I can understand your concern about unintentionally enabling CRC or the watchdog without latching in the data. Your method of already enabling the watchdog and CRC is an effective way to avoid this accidental condition. Alternatively, if it is possible to make sure that all devices on the avoid addresses 0x2B and 0x57 this will effectively avoid the issue as well.

    In what condition/EMI test did you see this behavior with the AD5420?

    As far as EMC/EMI test data, we have a TI Design that includes some extra protection circuitry along with the DAC8760. This design has been successfully tested against the IEC-61000-4 suite of transient tests. This will require some schematic and board layout modifications as several extra components are required to offer this level of protection. I have linked this design below. We now use the same protection circuitry on many of our devices.

    http://www.ti.com/lit/ug/tidu011/tidu011.pdf

    Thanks,

    Garrett

  • Hello,

    Just to piggy-back on Garrett's comments given some pretty extensive experience with this device...

    Concerning the material in Tables 2 and 4:

    The WDT topic should only be of concern for applications which implement the ALARM pin. Otherwise there are no impacts to any other functionality, and therefore the behavior would basically go unnoticed.

    Similar the CRC topic would only really be of concern for applications which implement daisy-chain as the only observable signature is a change in behavior for the SDO in daisy-chain. If this is not implemented then compliance with the CRC is not necessary. This one is particularly of importance to consider since, as indicated on page 25, the CRC calculation is performed every 32-clocks irrespective of the LATCH signal.
  • It would be convenient to run the DAC8750 as a superset of our existing AD5420 software (i.e. the same basic SPI communications, but additional registers). The unlatched commands break this possibility by for our multi-device SPI bus by forcing the Watchdog and CRC upon us when we don't need them.

    We have other SPI devices with CRC protection, so this is no big deal to us. Watchdog is inconvenient more than a problem-- We have 2 watchdogs in our system and don't need another one.

    In other words, you almost had an ideal solution with backward compatibility with AD5420 software and the additional DAC8750 features.

    Even less than ideal, we will be evaluating the DAC8750 in the near future.